UNDERSTANDING THE MEDICARE RADV AND ......and risk adjustment solutions and predictive models for...

3/17/2015

1

UNDERSTANDING THE MEDICARE RADV

AND MARKETPLACE IVA PROCESSES

AGENDA

• Who is Quadralytics?• Risk Adjustment 101• Medicare Risk Adjustment Data

Validation (RADV) Audits• Health Insurance Marketplace

Initial Validation Audits (IVA)• Understanding the Risk • Questions• Appendices

Page 2© 2014-2015 Quadralytics, LLC

OUR MISSION

Our mission is to provide accurate and timely consulting & analytic services to our healthcare partners to assist them in

formulating a comprehensive and unmitigated snapshot of members and providers based on

quality metrics, risk adjustment, and operational effectiveness.

© 2014-2015 Quadralytics, LLC Page 3

3/17/2015

2

QUADRALYTICS, LLCAbout Us

• Founded in 2011• Primary focus is providing

analytical solutions and consulting support to managed care organizations

• Team of consultants, SME, and technical staff with “hands on” experience

• Experience developing HEDISand risk adjustment solutions and predictive models for Medicare, Medicaid, and Health Exchange plans

Our Clients• Health plans• Physician organizations• Healthcare vendors• Self-Insured employers• Third-party administrators• ACO physicians


RISK ADJUSTMENT 101

WHAT IS RISK ADJUSTMENT?

• A method used to adjust bidding and payment based on the health status and demographic characteristics of an enrollee

• Pay appropriate and accurate reimbursement for subpopulations with significant cost differences

• Purpose: to pay plans accurately for the risk of the beneficiaries they enroll

• Why: access, quality, protect beneficiaries, reduce adverse selection, etc.


3/17/2015

3

TYPES OF RISK ADJUSTMENT• Prospective/Future Prediction:

• Uses historical diagnoses as a measure of health status and demographic information to predict future expense

• Data from 2014 used to predict expected costs in 2015

• Example: CMS Medicare HCC Model

• Concurrent (aka Retrospective):

• Uses historical diagnoses as a measure of health status and demographic information to predict expected expense for the current period done from a retrospective perspective

• Data from 2014 used to retroactively predict expected costs in 2014

• Example – HHS-CC model for the Health Insurance Marketplace


PROVIDER VIEW OF RISK ADJUSTMENT

My members are sicker

I documented the services

This is risk adjusted?

These numbers are not right


PAYOR VIEW OF RISK ADJUSTMENT

Why can’t they

document correctly?

Was that really a stroke in the office?

They need to hire a coder

Diabetes does not cure itself


3/17/2015

4

WHY DOES CMS CONDUCT AUDITS?

“To follow by faith alone is to follow blindly.”

- Benjamin Franklin


MEMBER EXAMPLE

• 60-year-old male• Originally disabled• Medicaid• Community• HCC 17 – Diabetes w/Acute

Complications• HCC 19 – Diabetes w/o

Complications• HCC 80 – Congestive Heart Failure• HCC 92 – Specific Heart

Arrhythmias• Interaction DM_CHF


HCC CALCULATION

Variable Accurate Missing

60‐year‐old male 0.411 0.411

Originally disabled 0.000 0.000

HCC 17 – Diabetes w/Acute Complications 0.339 0.000

HCC 19 – Diabetes w/o Complications 0.162 0.162

HCC 80 – Congestive Heart Failure 0.410 0.000

HCC 92 – Specific Heart Arrhythmia 0.293 0.293

Interaction for Diabetes and CHF 0.154 0.000

Total Hierarchical HCC weight 1.607 0.866

Annual payment (assumes $800/mo.) $15,427 $8,314

Payment Difference $7,113

Medical expense (85% MLR) $12,960 $12,960

Profit/Loss $2,467 ($4,646)


3/17/2015

5

MEDICARE AND RADV

MEDICARE HCC MODEL

• Model is prospective – previous diagnosis data used to predict future member expense

• Model is hierarchical – hierarchies apply to disease categories

• Model was essentially unchanged from 2004 implementation until 2014 payment year

• Risk scores correlate directly to plan payment


2013 VS. 2014 HCC MODEL


3/17/2015

6

MEDICARE HCC AUDIT

• Unlike other Medicare audits, the HCC audits do not have clear guidelines

• Whether a diagnosis is acceptable is often left to plan interpretation

• This may be different than what CMS determines to be acceptable

• Every plan must determine its acceptable level of risk

• Even when CMS provides guidelines, they are not always clear


ACCEPTABLE PROVIDER SPECIALTIES…


…OR ARE THEY?


3/17/2015

7

CMS RADV AUDIT PROCESS

• Plan is notified of RADV audit

• Roughly 600 Medicare contracts and only 30 plans are selected annually

• Odds of being selected for a RADV Audit: ~ 5% per year

• CMS selects 201 members for audit

• Three strata – low, medium and high risk scores

• Plan required to provide support for every HCC via medical record submission to CMS


ARE YOU AT RISK?

Signs your plan may be at risk for a RADV:

• Large change in year-over-year risk scores –CMS will focus on plans with big increases in score to ensure it is correct

• Very few delete records – if you are not doing deletes, you are not reviewing your own submissions for accuracy and correcting errors

• Other corrective actions – has your plan been reviewed for something else? It may increase your likelihood of audit as CMS sees you as a risk


All Possible Members

Members effective in claims year

All Year?

Was the member with you all year?

Current Year

Member still effective with plan 1/1 payment

year

ESRD

No ESRD Dx during 13 mo period

Hospice

Member not in hospice during 13 mo

period

Had Part B

Had Part B coverage for the data collection

period

Target Population

WHICH MEMBERS ARE INCLUDED?

Had an HCC

Diagnosis mapping to an HCC in claim year

© 2014-2015 Quadralytics, LLC

3/17/2015

8

CAN I REALLY SEND IN THAT MANY RECORDS?

• While the original RADV guidelines allowed for only the one “best medical record,” the new RADV guidelines have changed

• Plans can now submit up to five medical records to support a diagnosis and HCC

• The same medical record can be used to support multiple HCC for a member as well

• But the “best medical record” may not always be the best record to submit


HOW WILL I KNOW HOW THE PLAN DID?

CMS will issue a “Preliminary Audit Report of Findings” (AROF)

• Shows HCC-level validation and errors and eligibility for dispute

• At enrollee-level, AROF will show revised score and payment

• Information and instructions for Medical Record Dispute (MRD) will be included with report

• Plans allow to dispute findings only on certain types of RADV-related errors


PLAN HAS MULTIPLE LEVEL OF APPEALS

• Plans can file initial appeal via MRD process for review by “Hearing Officer”

• The plan must:

• File appeal within 30 days from receipt of AROF

• Submit the “One Best Medical Record” from records submitted to IVC for this review though it does not have to be the record audited


3/17/2015

9

PLAN HAS MULTIPLE LEVEL OF APPEALS

• Plan will receive “Audit Report Post Medical Record Review,” detailing results similar to AROF along with additional appeals instructions

• Only other appeal option is to CMS Administrator


ERROR EXTRAPOLATION

CMS Identifies HCC Errors

• Charts are read 2x by IVC• Plan notified of error

CMS Extrapolates Error

HCC 17

HCC 15

HCC 19

No HCC

• HCC 17 drops to HCC 19 .248 - .459 = (.211)

• Multiply By Benchmark $800 * (.211) = (168.80)

• Extrapolate to Population(168.80) * 8,000 = $1,350,400

• Other HCC for same member can change

• Interactions may no longer apply


INITIAL VALIDATION AUDIT (IVA)

3/17/2015

10

NOT YOUR MOTHER’S 3 “Rs”

THE 3 “Rs”

• The three keys to the risk adjustment and revenue of the Health Insurance Marketplace are:

• Risk Adjustment – the adjustment of payment based on the demographic factors and severity of the illness of the member

• Risk Corridors – The limiting or sharing of losses by the plan by HHS across all membership

• Reinsurance – The limiting of loss on an individual member basis


MEDICARE RADV VS. MARKETPLACE VALIDATION (IVA)

Item Medicare RADV Marketplace IVAPayment Years 2011 – Forward 2014 ‐ ForwardTimeline 2‐3 years after payment Six months after year‐endMinimum Plan Size Every Plan Not AddressedNumber of Plans Audited Approximately 30 AllMembers Stratified – 3 Strata Stratified – 10 StrataDiagnoses Included Thru 13 months after year‐end 4 months after year‐endMedical Records All Supporting All SupportingExtrapolation Applied to Strata Not Currently DefinedAppeal Process Defined DefinedFFS Offset Included – Est. 11% Not Applicable 2014/15Clarity Vague VagueFirst Round Audits Conducted By

CMSPlan Contracted

Vendor


3/17/2015

11

HHS-HCC MODEL

• More diagnoses are included and map to additional HCC because of broader disease implications for the commercial population

• What occurs in the year, affects payment for the year –retrospective or concurrent payment model

• Differences in plan type (Bronze, Silver, etc.) affect the risk score and associated payment

• Model is a zero-sum – if one plan’s risk score is higher than another plan’s, the lower risk score plan will have to make payments to higher risk score plan


ZERO SUM PAYMENTS

Plan Plan A Plan B Plan C

Initial Revenue $10,000,000 $10,000,000 $10,000,000

Initial Risk Score 1.15 1.07 1.23

Normalized Risk Score 1.00 0.93 1.07

Revised Revenue $10,000,000 $9,304,347 $10,695,653

Payment Change $0 ($695,653) $695,653


MODEL POPULATION

• Because the HHS Model includes a much more varied population than the Medicare model, some additional changes were necessary• Age groups include

infant through adults and seniors.

• Age groups are banded smaller for children and infants


3/17/2015

12

INITIAL VALIDATION AUDITUnlike Medicare Advantage, the Health Insurance Marketplace Initial Validation Auditors are contracted by the plan

• Both Health Insurance Marketplace and “Off-Exchange Plans” are included

• Members with and without HCC will be audited• All auditors must be certified by the American Association of

Professional Coders (AAPC) or the American Health Information Management Association (AHIMA)

• Senior auditors must have at least three years of experience in 2014 & 2015 and five years in 2016 and beyond

• Enrollment sources will be verified• Initial Validation Auditors must be free from conflicts of

interest


CONFLICTS OF INTEREST

• Issuer must attest to being conflict free to the best of its knowledge

• Neither the issuer nor any member of its management team (or any member of the immediate family of such a member) may have any material financial or ownership interest in the initial validation auditor

• Owners, directors and officers of the issuer may not be owners, directors or officers of the auditor (and vice versa)

• Audit Team members may not be married to, in domestic relationship with or immediate family of owners, directors, officers or employee of the issuer

• The initial validation auditor may not have had a role in establishing any relevant internal controls of the issuer related to the risk adjustment data validation process


AUDIT STRATA

No HCC – Demographic Only

Adult High Risk Score

Adult Medium Risk Score

Adult Low Risk Score

Child High Risk Score

Child Medium Risk Score

Child Low Risk Score

Infant High Risk Score

Infant Medium Risk Score

Infant Low Risk Score

80 % of Members

20% of Members


3/17/2015

13

MEMBERS WITH NO HCC

For enrollees without risk adjustment HCCs for whom the issuer has submitted a risk adjustment eligible claim or encounter, HHS would require the initial validation auditor to review all medical record documentation for those risk-adjustment eligible claims or encounters, as provided by the issuer, to determine if HCC diagnoses should be assigned for risk score calculation, provided that the documentation meets the requirements for the risk adjustment data validation audits.


ENROLLMENT VALIDATION

The initial validation auditor would validate information by reviewing plan source enrollment documentation, such as the 834 transaction, which is the HIPAA-standard form used for plan benefit enrollment and maintenance transactions. These enrollment transactions reflect the data the issuer captured for an enrollee’s age, name, sex, plan of enrollment, and enrollment periods in the plan.


ISSUER AUDIT RISK

• While no direct financial penalties will result from the 2014 and 2015 payment year audits, the possibility of financial penalties and further audit does exist:• Office of the Inspector General (OIG) – as noted in the

OIG Work Plan, the OIG is cracking down on over-coding of HCC.

• False Claims Act – knowingly submitting false diagnoses

• Whistleblowers – disgruntled employees, etc. may cry foul.


3/17/2015

14

UNDERSTANDING THE RISKS

BLIND FAITH

“Blind faith in your leaders or anything will get you killed.”

- Bruce Springsteen, “War”


BLIND FAITH

“Blind faith in your providers and claim submission will get you adverse

findings.”

- Scott Weiner, Quadralytics


3/17/2015

15

TOP 10 MEDICARE RISK ADJUSTMENTCODING ERRORS

1. The record does not contain a legible signature with credential.

2. The electronic health record (EHR) was unauthenticated (not electronically signed).

3. The highest degree of specificity was not assigned the most precise ICD-9-CM code to fully explain the narrative description of the symptom or diagnosis in the medical chart.

4. A discrepancy was found between the diagnosis codes being billed versus the actual written description in the medical record. If the record indicates depression, NOS (311 Depressive disorder, not elsewhere classified), but the diagnosis code written on the encounter document is major depression (296.20 Major depressive affective disorder, single episode, unspecified), these codes do not match; they map to a different HCC category. The diagnosis code and the description should mirror each other.


TOP 10 MEDICARE RISK ADJUSTMENTCODING ERRORS

5. Documentation does not indicate the diagnoses are being monitored, evaluated, assessed/addressed, or treated (MEAT).

6. Status of cancer is unclear. Treatment is not documented.

7. Chronic conditions, such as hepatitis or renal insufficiency, are not documented as chronic.

8. Lack of specificity (e.g., an unspecified arrhythmia is coded rather than the specific type of arrhythmia).

9. Chronic conditions or status codes aren’t documented in the medical record at least once per year.

10. A link or cause relationship is missing for a diabetic complication, or there is a failure to report a mandatory manifestation code.

http://news.aapc.com/index.php/2013/03/top-10-medicare-risk-adjustment-coding-errors/


UnsupportedDiagnoses

Supported DiagnosesNew

DiagnosesOriginal Claim Diagnoses

WHY DO MEDICAL RECORD REVIEW?Single Medical Record


3/17/2015

16

Original Claim DiagnosesUnsupportedDiagnoses


Diagnoses

WHY DO MEDICAL RECORD REVIEW?

Two to Three Medical Records


Original Claim Diagnoses

Unsupporte

dDiagn

oses


Diagnoses

WHY DO MEDICAL RECORD REVIEW?

Four or More Medical Records


CLAIMS DATA SUBMISSION

Advantages

• Chart review volume would be too great if we had to look at every record

• Can provide additional dates of services for a diagnosis beyond what is found via chart review

Disadvantages

• “75%” Accurate• Will not stand up to a RADV

Audit• Limited to how many the

provider can submit on a claim

• May not be able to tell if the service was done by an acceptable provider


3/17/2015

17

MEDICAL RECORD REVIEW

Advantages

• More accurate than claim submission only

• More complete than claim submission

• Able to identify the provider of service

• Additional diagnoses that may not have been on claim

• Fix the “30/30” issue

Disadvantages

• Time consuming• Intrusion on the provider

office• Retrospective • Chart coding is often open

to interpretation• Physician handwriting• EMR issues


PROSPECTIVE ASSESSMENTS

Advantages• Provides real-time picture of the

patient

• Provides a method to address care for home-bound or facility-bound patients

• Provides a look into the member’s living conditions

• More complete than the typical physician’s office health exam

• Not just about risk adjustment

• Provides complete and accurate documentation for RADV support depending on quality of data capture

Disadvantages

• More costly than office visit

• Office visit - $45-205 in Dallas

• Prospective Assessment ($300+)

• Physicians often see it as competition to their services

• Breaks the PCP/member relationship if not done correctly.

• Changes to CMS guidelines


PAPER VS. EMR RECORD

Paper

• Often not much more than a “super bill”

• Poor handwriting leads to misinterpretations

• Need legible signature and credentials on each page

• Need date on each page• Need member name on

each page

Electronic Record

• Usually cleaner than paper

• Menial tasks that must be done on a paper claim are done automatically.

• Several issues do exist with EMR records• Cloning• Drug lists not updated• Meaningful use


3/17/2015

18

REDUCING RISK

WHAT CAN BE DONE TODAY?

• Assess organizational readiness

• Assess data quality

• Validate existing charts

• Acquire and abstract charts where gaps exist


ASSESS THE ORGANIZATION• What does your Revenue Improvement

Program look like?

• RADV Response Team includes:• Business Sponsor (Senior Executive) • Medical Directors to call doctors• Executives to call office managers• Project Manager(s)• Review/Audit staff• Other Team Members

• Meet internally to develop strategy for RADV and determine need for assistance from vendor

• Are policies and procedures up-to-date?


3/17/2015

19

ASSESS DATA

Assess and clean up data• Have “Deletes” been processed

for bad data?• Code Sets

• Specialty codes (recently released)• CPT codes – may be acceptable

provider, but not face-to-face visit

• Are all RAPS (EDPS) resubmitted?

• Are specialty codes updated?• Are CPT/Dx codes reviewed?

• Update policies and procedures


CHARTING THE COURSE

• Which HCC do medical charts substantiate?

• Are the diagnoses from acceptable providers?

• Are “Rule-out” diagnoses used?

• What is the frequency of the diagnoses?

• If using a vendor, have all charts been reviewed?


TOP 10 COMPLIANCE ISSUES#3 Electronic Medical Records

Some early adopters of Electronic Medical Records (EMR) software are now having to respond to “cloning” and/or “carry over” concerns raised by ZPICs and Program Safe Guard Contractors (PSCs).

“These audits appear to be the result (at least in part) of inadequately designed software programs which generate progress notes and other types of medical records that do not adequately require the provider to document individualized observations. Instead, the information gathered is often sparse and similar for each of the patients treated.”

http://www.zpicaudit.com/2011/01/top-ten-health-care-compliance-risks-for-2011/ (emphasis added)


3/17/2015

20

QUESTIONS?


ASK US HOW WE CAN HELP

Scott WeinerEmail: [email protected]

Phone: (757) 716-7061

Cell: (757) 553-8985

http://www.quadralytics.com


RADV EXTRAPOLATIONAppendix

3/17/2015

21

THE IMPACT

Calculation Sample Strata 1 ‐ Hi Strata 2 ‐Mid Strata 3 ‐ LowModeled Payment $1,679,213 $1,164,902 $364,531 $149,779

Modeled Errors $218,256 $150,125 $43,392 $24,739

Extrapolated Payment $204,061,950 $141,561,722 $44,298,702 $18,201,526

Extrapolated Errors $26,522,990 $18,243,524 $5,273,099 $3,006,367

Standard Deviation $238,449 $333,095 $138,927 $107,118

Variance Estimates $17,223,049,893,090 $110,952,392,725 $19,300,754,327 $11,474,234,776

Standard Error $4,150,066

Lower Bound $22,372,924

FFS Adjuster (5%) $10,203,097

Final Amount Due $12,169,827 6%

The total HCC‐related payment made by CMS to the plan


THE IMPACT


Modeled Errors $218,256 $150,125 $43,392 $24,739







FFS Adjuster (5%) $10,203,097


The net effect of payment errors on the model assuming

~17% error rate.


THE IMPACT


Modeled Errors $218,256 $150,125 $43,392 $24,739







FFS Adjuster (5%) $10,203,097


Modeled payment for 201 enrollees in the sample and

expand to my entire population of ~24,000


3/17/2015

22

THE IMPACT


Modeled Errors $218,256 $150,125 $43,392 $24,739







FFS Adjuster (5%) $10,203,097


Modeled payment errors for 201 enrollees in the sample and expanded to my entire population of ~24,000


THE IMPACT


Modeled Errors $218,256 $150,125 $43,392 $24,739







FFS Adjuster (5%) $10,203,097



THE IMPACT


Modeled Errors $218,256 $150,125 $43,392 $24,739







FFS Adjuster (5%) $10,203,097


(~121*110,952,392,725)+…


3/17/2015

23

THE IMPACT


Modeled Errors $218,256 $150,125 $43,392 $24,739







FFS Adjuster (5%) $10,203,097


The maximum pay back to CMS for overpayments = Extrapolated

Errors – Standard Error


THE IMPACT


Modeled Errors $218,256 $150,125 $43,392 $24,739







FFS Adjuster (5%) $10,203,097


The maximum pay back to CMS for overpayments = Extrapolated

Errors – Standard Error

For illustrative purposes only based on 5% of Extrapolated

HCC Payments


The Three “Rs” of HHS RiskAppendix

3/17/2015

24

MARKETPLACE 3 “Rs”


RISK CORRIDORS

• Similar to Part D plans at start-up; the federal government will apply risk corridors to profit and loss of individual health plans in- and out-of the Marketplace


RISK CORRIDOR – LOSS

• Plan has $125M revenue

• Plan expense ratio 15%

• Actual plan medical spend - $120M


3/17/2015

25

RISK CORRIDOR – GAIN

• Plan has $125M revenue

• Plan expense ratio 15%

• Actual plan medical spend - $10M


REINSURANCE

• Reinsurance designed to protect plans from impact of a few high risk member/catastrophic claims• For 2014, members with total claims in excess of $45,000

(attachment point) will be covered at 80% to a $250,000 maximum per member/claim

• For 2015, the attachment point is $70,000• Payments are funded from payment - all plans pay

whether they are in the exchange or not.• 2014 - $63 per member payment• 2015 - $44 per member

• Plans will typically carry traditional reinsurance above the $250,000 threshold.


UNDERSTANDING THE MEDICARE RADV AND ......and risk adjustment solutions and predictive models for...

Documents

Transcript of UNDERSTANDING THE MEDICARE RADV AND ......and risk adjustment solutions and predictive models for...